Mental health online forum users
Conditions
Interventions
Group 1: Single-arm, non-interventional secondary data analysis of publicly accessible and anonymous online forum posts. No study participants are recruited, interviewed, or treated. Approximately 80,
Sponsors
Charité - Universitätsmedizin Berlin
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: In this study, no individuals are recruited and no study participants are actively involved. Recruitment therefore refers to the process of secondary data collection. This process is based exclusively on previously published, publicly accessible text posts from predefined (sub)Reddit forums. There is no interaction with individuals in the forums. Users are neither contacted, observed, surveyed, nor otherwise involved in the study. Thus, the analysis refers exclusively to the qualitative text entries. For each included post, only the year of collection, the post title, and the post content are extracted. In addition, a random post ID is assigned locally. Direct identifiers, usernames, profile data, comment histories, or any other person-related contextual information are explicitly not collected. Therefore, multiple posts are also not linked across individuals or over time. Inclusion criteria - The text post originates from predefined, publicly accessible Reddit forums with a clear connection to mental health topics. - The post was published within the study period from 01.01.2010 to 01.08.2026. - The post is text-based and contains at least 60 words. - The post contains an explicit description of experiences, feelings, distress, coping, help-seeking, or everyday experiences in connection with psychological distress or mental health conditions. - The post is publicly accessible at the time of data collection and has not been deleted, removed, or made inaccessible by the platform.
Exclusion criteria
Exclusion criteria: The following text posts will not be included: - Posts that consist exclusively or predominantly of images, videos, audio content, GIFs, memes, screenshots, surveys, or other non-text-based content. - Posts that consist exclusively or predominantly of web links, quotations, news articles, scientific sources, documents, or other external sources. - Posts without sufficient relevance to subjective experience in the context of psychological distress or mental health conditions. - Posts that primarily contain advertising, spam, recruitment calls, surveys, petitions, commercial content, or content that appears automated.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| The primary endpoint is the structural correspondence between the symptom structure reconstructed in a data-driven manner from the forum posts and the HiTOP model. The analysis will examine whether symptoms extracted bottom-up from qualitative, text-based data yield a hierarchical structure consisting of components, syndromes, and higher-order spectra that corresponds to the central HiTOP domains, such as Internalizing, Externalizing, Detachment, Thought Disorder, and Somatoform Distress. The analysis will be based on aggregated binary or categorical symptom variables. Within a “multiverse” analytic approach, the structural correspondence will be examined using, among other methods, tetrachoric and polychoric correlations, Exploratory Graph Analysis, hierarchical procedures such as Bass-Ackwards analysis and the Schmid-Leiman transformation, as well as categorical fit indices, explained variance, centrality measures, and reliability or stability metrics. | — |
Secondary
| Measure | Time frame |
|---|---|
| 1. Analysis of the cognitive and symptom-level network structure using mixed graphical models, Ising networks, and Bayesian networks. 2. Prediction of specific clinically relevant symptoms (such as depressive symptoms and suicidality) 3. Description and explanation of the network structure of reported loneliness and social isolation and its central influencing variables 4. Analysis of associations between expressed resources or coping strategies and central reported symptoms 5. Assessment of the stability and coverage of individual HiTOP spectra 6. Assessing associations between expressed cognitive biases and overall symptom burden | — |
Countries
Germany
Contacts
Public ContactDrin Ferizaj
Charité - Universitätsmedizin Berlin
Outcome results
None listed